1. Field of the Invention
The present invention relates in general to the field of energy recovery systems, and in particular, to systems, program product, and methods related to energy efficiency assessments of single and multiple utilities, grass-roots new and existing heat exchanger network designs, and current and future retrofit modifications.
2. Description of the Related Art
Many different types of processes consume multiple steam levels and electricity to obtain an output result, or to produce a required product or compound. For large-scale processes that, for example, consume significant amounts of fuel and steam, it is preferable to optimize the consumption of energy through careful operation, design or reconfiguration of the plant and the equipment used. Further, in some industrial manufacturing processes, specific streams of material flows need to be supplied to different types of equipment and machinery at specific temperatures. These material flows may need to be heated or cooled from an original starting or supply temperature to a target temperature. This, in turn, requires the consumption of steam to heat specific streams and consumption of water, for example, to cool down specific streams.
The total energy employed or consumed by the industrial manufacturing processes can be optimized to a global minimal level, for example, through careful placement and configuration of specific material streams with respect to one another. There may be, for example, the potential for hot streams that require cooling to be placed in proximity with cold streams that require heating. Streams having thermal energy already present that need to be removed (waste heat) or streams that need to have heat added can be associated with one another to optimize the energy consumption of the process. A network of heat exchangers (HENs) can be synthesized to provide a medium for utilizing this waste heat to provide heat to those streams that need to have heat added. This heat exchanger network can be a very important sub-system in any new plant.
HEN designs face a high level of operational changes along their lifetime. These changes can be short-term, such as, for example, as a result of process disturbances and/or uncertainty in feed stock conditions and product demand, and can be long-term such as, for example, as a result of the need to process more raw materials that warrants the debottlenecking of components of the facility including the HEN to increase its capacity. Nowadays, and since late seventies of the last century, another important long-term factor has been the continuous escalation in energy prices in a rate that is higher than the rate of increase of plant energy system equipment costs. This relatively continuous escalation in energy costs has prompted a need to periodically modify the facility's HEN to increase the facility's HEN waste energy (quantity and quality) recovery along the lifetime of the facility which can reach 30 or more years.
In such cases, the HEN retrofit objective/task is to produce a practically implementable cost effective HEN design modification that satisfies the new process objective and its new operating constraints. The inventor has recognized that there are many possible modifications for an existing HEN to retrofit the original design to the new objective. They can include one or more of the following: process operating and design conditions modifications, existing HEN topological/structural modifications, and existing HEN unit design modifications and parametric modifications including heat transfer enhancements to improve heat transfer (U), etc.
Theoretically such problem can be solved using optimization, where we have an objective to minimize total cost of the existing HEN modifications and the rest of the plant's process units design and operation modifications minus the realized energy cost saving. Accordingly, the inventor has recognized that it will be very beneficial to the world's waste heat recovery efforts to have systematic process steps that use or employ all possible degrees of freedom identified above to enhance waste heat recovery of any industrial facility, which can be automatically converted to a percentage of debottlenecking in the plant's energy system.
Also recognized by the inventor is that it will also be very beneficial to any industrial facility to have systematic process steps that can use or employ all possible combinations of available degrees of freedom to improve the current efficiency of any facility's HEN now, two years from now, five years from now and so on till the end of the plant lifetime without compromising any retrofit effort done at earlier time. In other words, there is a need for a cost effective way for plant owners to be able to assess and retrofit their HEN at today's energy prices to obtain present energy cost savings, and to do so in such a manner that when decision-makers want to assess and perform another retrofit to save on energy costs, two to three years later due to another increase in energy prices, they can do so without contradicting the actions done in the previous retrofit projects.
Accordingly, recognized is the need for systematic methods, systems, and program product that enable industrial facilities to have the ability to continue modifying their HEN, yet not be handicapped with a situation in which they have no options or only very expensive/unrealistic and/or contradictory retrofit options to be able to capture more waste heat in their production facilities.
Further, recognized is the need for an abstract process that can test the results of any other retrofit method existing today or in the future including the results of mathematical programming models and that can assess its results from both a number of units added and waste energy recovered perspective and to determine whether the HEN has a healthy topology that can allow more retrofit projects in the future. In other words, there is a need for a new systematic method, system, and program product for retrofit-now-with-future-retrofits-in-sight that renders structures capable of continual plant lifetime retrofit for waste heat recovery purposes.
Conventional methods for HEN retrofit use either: the pinch technique and its modifications, the network pinch technique and its enhancements, or mathematical programming techniques. The methods of HEN retrofit that use the pinch design method and its modifications consists of an ad hoc targeting stage followed by a design modification stage. In the ad hoc targeting stage a global Δt_min is assumed and used to set an energy target, and the network is then drawn using a grid diagram. The relative positions of the heat exchangers to an “assumed to be fixed” pinch point resulting from the assumed global Δt_min value is then examined to identify HE units that exchange heat across this “factiously fixed” pinch point (location). In the design modification phase, the heat exchangers and their associated heaters/coolers are relocated semi-systematically/manually relative to the factiously fixed pinch location so that they no longer exchange heat across the location of the factiously fixed pinch point.
In the conventional heuristics-based HEN retrofit methods using pinch technology alone or in combination with mathematical programming model, a pay-back period for a “onetime retrofit” project without future retrofit projects in mind/sight, e.g., two years, is first selected. Next, the required energy to be saved and new heating or cooling loads is determined from composite curves using a global ΔT_min that coincides with the location of the process pinch point. The existing network is drawn on what is referred to as a grid diagram where the units working across the process pinch temperature are highlighted. The designer then uses his/her experience to correct the situation of violating the process pinch temperature, relying on an assumption that such changes will not change the pinch location.
Being that the starting of problem/initialization is based upon heuristics, it proceeds without future retrofit projects in mind/sight to a retrofit solution. Using a different starting point, however, will likely result in a completely different network structure retrofit solution. Hence, as many alternative designs of similar performance are possible, it is mandatory to start the retrofit project with the so called “right” initialization global ΔT_min value. It is known to one of ordinary skill in the art experienced in the field, that different values of ΔT_min will produce different pinch locations which will lead to different network structures. As such, if one starts with the wrong ΔT_min, the wrong structure will be obtained, and evolutionary heuristics and optimization techniques (NLP) will not lead to the best results since radical moves/changes from one structure to another are not feasible or possible. While this retrofit solution finding is based upon the first law of thermodynamics, it is not conducted in a systematic way to reach the originally defined targets, rather it is an all-at-once-approach (one-time retrofit package). That is, if the designer cannot accept the entire solution structure for one reason or another (e.g., capital needed and/or time for implementation, affect on other process units, etc.) the package will have to be rejected. This can be troublesome as the targets defined for area and investment costs are not rigorous and very crude.
The methods of HEN retrofit that use the network pinch based methods employ procedures to determine the maximum possible waste heat recovery in the network via exploitation of only the network utility path(s)—e.g., through the addition of more HE unit surface area. It is an evolutionary method that repeats the pinching of the network via iteratively minimizing ΔT_min between streams in each process-to a process HE unit to zero, to create what is known as a new “assumed fixed” network pinch that does not change its location with process and/or structural modifications. Then using same rules of the pinch technology, HE units that exchange heat across the new factiously fixed network pinch are relocated. Commercial software are currently using such methods complemented by mathematical programming techniques, in form of LP, NLP and MILP to select streams matches, optimize branches flows in the base design networks, and/or help find minimum cost topology modifications to the existing network.
In the network pinch method that uses a mix of mathematical programming models and the network pinch concept, the existing network streams are first forced to pinch the cold composite curve using composite curve or mathematical programming models to locate the so called network pinch. The method then uses the pinch technology rules to find topological modifications such as re-sequencing, introduction of new heat exchangers, streams splits, etc., followed by or interlaced with the employment of a network optimization model for loads/branches flow using NLP techniques. The use of pinch technology rules to find topological modifications in the existing HENs to overcome the so called network pinch, however, is not rigorous since it is based upon the assumption that the network pinch and the process pinch are not going to change its location due to the taken topological modification. As the assumption is not correct, topological modifications can have counter effects on each other to the extent that they may completely negate the possibility of doing future retrofit projects. That is, it may become impossible to employ a certain retrofit project due its counter effect on the previous projects, and so on, making the retrofit projects in sustainable.
For example, if the current HEN retrofit process is conducted using pinch technology concept including the evolutionary network pinch method, and then if in at least one time at any future network retrofit process a different pinch point location arises, e.g., due to process parametric and/or structural modifications, and if the different pinch point location indicates a need for a lesser minimum number of units than the one used in previous HEN retrofit project, the result may be a wasting of some of the HE units already added in the previous HEN retrofit.
Accordingly, recognized by the inventor is that it would be beneficial to the facility to have a process whereby if the pinch point for a retrofit project changes its location due to the current and future network retrofit project solutions (including changes to process parametric and structural modifications), such situation would not detrimentally affect the network's topology retrofit. That is, there would be no significant effect on streams matching and/or change in minimum number of units needs previously completed or to be completed in the future, with the exception of the need for enhancing the current network UA as part of the upcoming network retrofit project and the future ones.
Acceptance of the pinch technology method and its modifications has been attributed to the interaction that exists between the designer and the evolving retrofit solution. Such existence of the designer in-the-loop enables him/her to perform an in-depth cost estimation, accepting or rejecting a solution aspect as desired. It also provides a crude targeting phase which can help inform plant managers about the merits of conducting HEN retrofit project before starting the retrofit exercise, and it can accommodate many degrees of freedom for the retrofit solution, which are generally too expensive to be practically considered in the mathematical programming approach (e.g., using the MINLP model) especially for an industrial size problem.
Acceptance of the network pinch method, adopted in commercial software such as SPRINT, has been attributed to its ability to enable the user to search for the possible retrofit action automatically to determine the required topology modification. Further, using mathematical programming models, the software employing the network pinch method conducts an optimization stage for fixed topology to exploit utility path(s), best load allocation on existing units, and branches flow distribution. It employs an evolutionary approach to retrofit the HEN using several mathematical programming models, and the network pinch concept to guide the search for retrofit solutions. Such use, however, is crude, based upon heuristics, and is not rigorous (repeatably accurate).
Although, as noted previously, most commercial software are currently using mathematical programming techniques to complement other methodologies, pure mathematical programming techniques have been less well received, particularly in the industry community. Mathematical programming techniques are typically conducted in one stage using MINLP or in two stages using MILP followed by NLP using a superstructure that embeds “all possible solutions” with a cost-based objective function consisting of cost correlations to identify the retrofit solution. Mathematical programming techniques use a deterministic search method and heuristic/approximating search methods employing, for example, Genetic Algorithms, Tabu search, simulated annealing algorithms and hybrid methods to find the global solution of the retrofit problem. It is expected that if accurately embedded in the problem superstructure, the mathematical programming technique can provide for the simultaneous consideration of “all possible modifications.”
Such methods above are referenced in the following papers: Tjaan N. Tjoe and Bodo Linnhoff, “Using pinch technology for process retrofit”, Chemical engineering, pp 47-60, April (1986); N. Asante and X. X. Zhu, “An automated and interactive approach for heat exchanger network retrofit”, Transactions of Institution of Chemical Engineers, vol. 75, part a, pp. 349-360, (1997); Amy Ciric and Christodoulos Floudas, “A mixed integer non-linear programming model for retrofitting heat exchanger networks”, Industrial & Engineering Chemistry Research, vol. 29, pp. 239-251, (1990); T. Gundersen and L. Naess, “The synthesis of cost optimal heat exchanger networks”, Computers and Chemical Engineering, vol. 12, pp. 503-530 (1988); Kevin C. Furman and Nikolaos v. Sahinidi, “A critical review and annotated bibliography for heat exchanger network synthesis in the 20th century”, Industrial & Engineering Chemistry Research, vol. 41, pp. 2335-2370 (2002); R. Smith, M. Jobson and I. Chen, “Recent development in the retrofit of heat exchanger networks”, Chemical Engineering Transactions, vol. 18, pp 27-32 (2009); E. Rezaei and S. Shafiei, “Heat exchanger networks retrofit by coupling genetic algorithm with NLP and ILP methods”, computers and chemical engineering, vol. 33, pp 1451-1459, 2009; and R. Smith, M. Jobson and I. Chen, “Recent development in the retrofit of heat exchanger networks”, Applied Thermal Engineering, vol. 30, pp 2281-2289 (2010).
Each of the above described existing methods have some significant disadvantages. For example, in the most widely used pinch technology method and its modifications, the method is non-systematic and involves heuristics that produce solutions whose quality depends entirely on the designer experience in applying such heuristics—i.e., the global ΔT_min should be used in determining the location of the pinch for an assumed best global ΔT_min. However, as the value of assumed best global ΔT_min changes, the location of the pinch point changes, and therefore, the assessment of which HE units transfer heat across the pinch also changes, making the proper selection critical. The same also applies with respect to process changes that result in the same output. Further, retrofit is provided as a one package solution that has no consideration for step-by-step modifications and practicality for future retrofit projects. Employment of a heat exchanger(s) having a larger surface area during a later retrofit to reduce ΔT_min can result in a network having one or more heat transfers which cross the pinch.
Accordingly, the inventor has recognized that such methodology is not systematic, requiring iteration, and that pinch method initialization using an assumed best global ΔT_min is ad hoc, and thus, such methodology could not be used as a standard method for plant-lifetime retrofit. The inventor has also recognized, through analysis of designs provided by the network pinch method, for example, that the network pinch method can also result in a network having one or more heat transfers which cross the pinch during a later retrofit. This is because the method incorrectly assumes that changes in topology of a process HEN will not affect the process pinch and that any topology change which only creates an opportunity to load the network pinch will not increase the R_max. In fact, through detailed analysis, the inventor has found that both parameters can change. Accordingly, such methodology could also not be used as a standard method for plant-lifetime retrofit.
The mathematical programming-based methods also have significant disadvantages. For example, the mathematical programming-based methods provide little or almost no scope for user interaction. Although it is currently possible to solve larger problems than before utilizing such methods, the current capacity is still not to the scale needed in many industrial applications. The problem of the HEN retrofit is mathematically considered to be NP-hard. As such, the inventor has recognized that including all possible process design modifications, different types and configurations of HE units and so on in the superstructure is impractical and will make the problem harder, intractable and impossible to solve.
In addition, the MINLP model cost objective function needs comprehensive data for the model solution to be realistic. The required information/data, however, is normally not available during the conceptual phase of the retrofit project, especially in a typical de-centralized engineering environment where projects cost control depends entirely on rigorous cost calculations, which typically, if not always, are made after the retrofit project basic engineering, and which are carried out at different departments, at least in most of the companies. Further, it would be neither practical nor logical to assume with high fidelity all possible retrofit solutions that could potentially be implemented for a given plot plan of a facility or to rigorously develop cost correlations to each before running the mathematical program—especially when there is always more than one option to carry out a designated retrofit modification. As such, the inventor has recognized that mathematical programming techniques will often lead to designs which include heat exchange across the pinch, which would be almost impossible to correct in future retrofits to recover extra heat.
Accordingly, recognized by the inventor is that neither one of these methods (pinch technology, network pinch and mathematical programming) can rigorously assess the existing network; rigorously target for the solution before the go ahead of the retrofit project for practically attainable waste heat recovery and/or minimum number of units that need to be added to reach desired energy consumption targets or part of it, and to systematically render solutions that can be implemented in phases with a guarantee that the solution implemented today (in phase one) will not become an obstacle, or need to be demolished, for a tomorrow retrofit solution to be implemented at later time in the future in phases two, three and so on. Also recognized is that the methods existing today do not employ all possible combinations of process changes that not only include the HEN, but also the rest of the plant design and operating conditions, for finding the best HEN retrofit solutions. Correspondingly, recognized is the need for methods, systems, and program product for managing heat exchanger network energy efficiency and retrofits for an industrial facility, which provide the above features.